Transglutaminase-mediated fibronectin multimerization in lung endothelial matrix in response to TNF-alpha.

Exposure of lung endothelial monolayers to tumor necrosis factor (TNF)-alpha causes a rearrangement of the fibrillar fibronectin (FN) extracellular matrix and an increase in protein permeability. Using calf pulmonary artery endothelial cell layers, we determined whether these changes were mediated by FN multimerization due to enhanced transglutaminase activity after TNF-alpha (200 U/ml) for 18 h. Western blot analysis indicated that TNF-alpha decreased the amount of monomeric FN detected under reducing conditions. Analysis of (125)I-FN incorporation into the extracellular matrix confirmed a twofold increase in high molecular mass (HMW) FN multimers stable under reducing conditions (P < 0.05). Enhanced formation of such HMW FN multimers was associated with increased cell surface transglutaminase activity (P < 0.05). Calf pulmonary artery endothelial cells pretreated with TNF-alpha also formed nonreducible HMW multimers of FN when layered on surfaces precoated with FN. Inhibitors of transglutaminase blocked the TNF-alpha-induced formation of nonreducible HMW multimers of FN but did not prevent either disruption of the FN matrix or the increase in monolayer permeability. Thus increased cell surface transglutaminase after TNF-alpha exposure initiates the enhanced formation of nonreducible HMW FN multimers but did not cause either the disruption of the FN matrix or the increase in endothelial monolayer permeability.

Increased recycling of (alpha)5(beta)1 integrins by lung endothelial cells in response to tumor necrosis factor.

Tumor necrosis factor (alpha) (TNF-(alpha) can change the interaction of lung endothelial cell monolayers with their extracellular matrix in association with an increase in endothelial monolayer protein permeability. Using immunofluorescence microscopy and flow cytometry, we determined if exposure of calf pulmonary artery endothelial monolayers to TNF-(alpha) may influence cell-matrix interactions by altering the clustering as well as internalization of the (&agr;)5(beta)1 integrins (or fibronectin receptors) on the surface of endothelial cells. Immunofluorescence microscopy revealed that TNF-(alpha) caused an increase in the intracellular staining of (alpha)5(alpha)1 integrins within structures similar to endocytic vesicles as well as an increase in antibody-induced clustering of the integrins at the cell periphery. Flow cytometric analysis of endothelial cells incubated at 37 degrees C after antibody-labeling of their surface (alpha)5(beta)1 integrins at 4 degrees C confirmed an increase in the rate of (alpha)5(beta)1 integrin internalization which was at least 3 times greater after TNF-(&agr;) exposure, based on the half-life for antibody-labeled surface integrins to reach equilibrium with non-labeled integrins within the intracellular pool. Interestingly, the total cell surface expression of (alpha)5(beta)1 integrins was relatively constant after TNF-(alpha) exposure despite the enhanced rate of internalization, suggesting an accelerated recycling of the internalized (alpha)5(beta)1 integrins back to the cell surface. This response was confirmed by the measurement of labeled integrin recycling, which showed a significant (P<0.01) increase in the rate of recycling of the internalized integrins in TNF-treated endothelial cells. Enhanced internalization and subsequent recycling of (alpha)5(beta)1 integrins by endothelial monolayers exposed to TNF-(alpha) may facilitate the redistribution of cell-surface integrins in response to this inflammatory cytokine and may also modify cell-matrix interactions leading to reduced integrity and increased protein permeability of the lung endothelial monolayers.

Lung matrix incorporation of plasma fibronectin reduces vascular permeability in postsurgical bacteremia.

Plasma fibronectin (pFN) can incorporate into the lung extracellular matrix (ECM) as well as enhance hepatic cell phagocytic removal of bloodborne microparticulate debris that can contribute to lung vascular injury. Treatment of human pFN (hFN) with N-ethylmaleimide (NEM) blocks its ECM incorporation but not its ability to augment phagocytosis. Using hFN purified from fresh human plasma cryoprecipitate, we compared the effect of NEM-treated hFN versus normal hFN on lung transvascular protein clearance (TVPC) in postoperative bacteremic sheep to determine whether the ability of hFN to attenuate the increase in lung endothelial permeability required its ECM incorporation. Sheep with lung lymph fistulas were infused with a sublethal dose of Pseudomonas aeruginosa (5 x 10(8)) 48 h after surgery. In the first study, sheep received either FN-rich human cryoprecipitate, FN-deficient cryoprecipitate, FN purified from cryoprecipitate (hFN), FN-deficient cryoprecipitate reconstituted with purified hFN, or the sterile saline diluent. In the second study, sheep received either 200 mg of purified hFN (group I), 200 mg of NEM-treated hFN (group II), or the saline diluent (group III). In the first study, the increase in TVPC after bacterial challenge was attenuated by FN-rich cryoprecipitate, hFN, or reconstituted FN-deficient cryoprecipitate (P < 0.05) but not by saline and FN-deficient cryoprecipitate. In the second study, TVPC increased by 2 h (P < 0.05) and peaked over 4-8 h (P < 0.05) at 380-420% above baseline in postoperative bacteremic sheep given the diluent (group III). In contrast, intravenous infusion of hFN, but not of NEM-treated hFN, significantly (P < 0.05) attenuated this increase of lung protein clearance. Thus the ability for the intravenously infused purified pFN to attenuate the increase in lung endothelial protein permeability in sheep during postsurgical bacteremia appears to require its ECM incorporation into the interstitial ECM of the lung.

TNF-alpha-induced matrix Fn disruption and decreased endothelial integrity are independent of Fn proteolysis.

Exposure of confluent pulmonary arterial endothelial monolayers to tumor necrosis factor (TNF)-alpha causes both a reorganization and/or disruption of fibronectin (Fn) in the extracellular matrix and an increase in transendothelial protein permeability. However, the factors initiating this response to TNF-alpha have not been defined. Because TNF-alpha can induce proteinase expression in endothelial cells, we determined whether proteinases cause both the alteration of the Fn matrix and the permeability increase as is often speculated. Incubation of calf pulmonary arterial endothelial monolayers with TNF-alpha (200 U/ml) for 18 h caused a disruption of the Fn matrix and an increase in transendothelial protein permeability. A reduced colocalization of cell-surface alpha5beta1-Fn integrins with the Fn fibers in focal contacts was also observed. TNF-alpha treatment of endothelial monolayers with matrices prelabeled with 125I-human Fn (hFn) did not cause the release of Fn fragments or alter the content of Fn antigen in the medium as analyzed by SDS-PAGE coupled with autoradiography. Both the content and fragmentation pattern of Fn within the cell layer and the insoluble Fn matrix also appeared unchanged after TNF-alpha exposure as confirmed by Western immunoblot. Fn-substrate zymography revealed that TNF-alpha increased the expression of two proteinases within the conditioned medium in which activity could be blocked by aprotinin but not by EDTA, 1,10-phenanthroline, leupeptin, or pepstatin. However, inhibition of the Fn proteolytic activity of these two serine proteinases did not prevent either the TNF-alpha-induced disruption of the Fn matrix or the increase in permeability. Thus the reorganization and/or disruption of the Fn matrix and the temporally associated increase in endothelial permeability caused by TNF-alpha appear not to be due to proteolytic degradation of Fn within the extracellular matrix. In contrast, decreased alpha5beta1-Fn integrin interaction with Fn fibers in the matrix may be important in the response to TNF-alpha exposure.

Lung matrix deposition of normal and alkylated plasma fibronectin: response to postsurgical sepsis.

Plasma fibronectin (Fn) can both enhance phagocytic clearance of microparticulate debris by macrophages as well as incorporate it into the lung extracellular matrix (ECM). The goal of this study was to document that N-ethylmaleimide (NEM)-treated human plasma Fn (HFn) would lose its ability to incorporate into the lung ECM in vivo even though it would retain its ability to stimulate test particle phagocytosis and bind to fibrin. Using dual-label immunofluorescence, we compared the lung deposition of purified normal HFn and NEM-alkylated HFn (NEM-HFn) after their intravenous injection into postoperative nonbacteremic and bacteremic sheep in relationship to the localization of endogenous sheep Fn. Two days after a sterile surgical thoracotomy, sheep were infused with either 5 x 10(8) Pseudomonas aeruginosa (postsurgical bacteremic model) or the diluent (nonbacteremic model). They also received a bolus 100-mg injection (5 min) of either HFn or NEM-HFn. Analysis of serial lung biopsies harvested at 2-h intervals demonstrated little deposition of NEM-HFn compared with HFn in the lung interstitial matrix of postoperative nonbacteremic sheep. In contrast, enhanced deposition of both HFn and NEM-HFn was observed in the lungs of postoperative bacteremic sheep. However, in the lungs of bacteremic sheep, HFn displayed a diffuse fibrillar deposition pattern in the lung characteristic of ECM incorporation, whereas the enhanced NEM-HFn deposition, especially in the interstitial ECM region of the lung, was primarily focal and punctate, with very little fibrillar incorporation. Immunofluorescent analysis with antibodies specific to fibrinogen, Fn, and lung macrophage surface antigens coupled with immunoperoxidase staining for HFn antigen revealed that the punctate fluorescence pattern was due to both the binding of HFn to fibrin and its colocalization with inflammatory cells. Thus treatment of plasma Fn with low concentrations of NEM will limit its normal in vivo fibrillar incorporation into the interstitial ECM region of the lung.

Reduced in vivo plasma fibronectin content of lung matrix during postoperative sepsis.

Sepsis after surgery, trauma, or burn contributes to altered lung endothelial permeability and respiratory failure. Fibronectin (Fn), an opsonic and adhesive glycoprotein, exists in both a soluble form in plasma and an insoluble form in the extracellular matrix (ECM). Recent studies [E. M. Wheatley, P. J. McKeown-Longo, P. A. Vincent, and T. M. Saba, Am. J. Physiol. 265 (Lung Cell. Mol. Physiol. 9): L148-L157, 1993] suggest that the ECM content of Fn may influence lung vascular permeability. We evaluated the incorporation of plasma-derived Fn (pFn) into the ECM of the lung during postoperative sepsis. Postoperative nonseptic and postoperative septic rats were compared, using a model of laparotomy followed by cecal ligation and puncture. To label the pFn pool, rats received intravenously 3 micrograms of purified rat 125I-labeled Fn/100 g body weight 6 h after surgery (laparotomy). 125I-Fn in the deoxycholate detergent-insoluble fraction of tissues was used to quantify matrix-incorporated Fn at 4 h after infusion with 125I-Fn. Septic rats exhibited a peripheral leukopenia as well as reduction in plasma volume, Fn halflife, and total pFn pool. Incorporation of pFn in the liver and spleen of postsurgical septic rats was not different (P > 0.05) from sham-operated (postsurgical nonseptic) rats, but incorporation was significantly decreased (P < 0.05) in vivo in the lung. However, under controlled in vitro conditions, lung tissue harvested from septic or sham-operated rats demonstrated a similar tissue incorporation of soluble 125I-pFn as well as similar rates of retention/turnover of ECM 125I-Fn, based on pulse-chase experiments. These data suggest that the in vivo inflammatory environment in the lung during postoperative sepsis, which cannot be reproduced in vitro, may alter the Fn content of the ECM of the lung. Such reduced levels of pFn in the lung ECM may be a factor influencing lung vascular integrity during postoperative sepsis.

Fibronectin attenuates increased endothelial monolayer permeability after RGD peptide, anti-alpha 5 beta 1, or TNF-alpha exposure.

Endothelial permeability can be altered by tumor necrosis factor-alpha (TNF-alpha), a cytokine released in association with inflammation-induced tissue injury. In the subendothelial matrix, fibronectin (Fn) influences endothelial cell adhesion by the interaction of integrins with RGD and non-RGD attachment sites in Fn. We compared the effect of TNF-alpha, RGD-containing peptides (GRGDSP), or antibody to alpha 5 beta 1-integrins on the protein permeability of bovine lung endothelial monolayers as assessed by transendothelial 125I-labeled albumin clearance. We also examined the influence of purified human plasma fibronectin (hFn) on this permeability response. TNF-alpha, RGD peptides, and antibodies to alpha 5 beta 1-integrins elicited a dose- and time-dependent increase in protein permeability as well as a reorganization and/or disruption of the endogenous Fn matrix. A control RGE peptide (GRGESP) as well as immunoglobulin G purified from nonimmune rabbit serum did not increase endothelial protein permeability or disrupt the endogenous fibrillar Fn pattern in the matrix. Likewise, a LDV peptide derived from the alternatively spliced type III connecting segment (IIICS) within bovine Fn (bFn) was unable to increase permeability of the bovine endothelial monolayer. Co-incubation of purified soluble hFn (300 or 600 micrograms/ml) with either TNF-alpha, the RGD peptide, or the antibody to alpha 5 beta 1-integrins prevented the increase in endothelial permeability. This protective effect was also observed when the purified hFn (600 micrograms/ml) was added after the TNF-alpha-induced increase in endothelial permeability had taken place. Immunofluorescent analysis confirmed the incorporation of the hFn into the subendothelial matrix and its co-localization with the endogenous bFn. The similar alteration of the subendothelial matrix after exposure to RGD peptides, anti-alpha 5 beta 1-antibodies, or TNF-alpha, coupled with the ability for hFn to attenuate the permeability increase typically elicited by all three agents, suggests that disruption of cell-matrix interactions may be the mechanism by which TNF-alpha alters endothelial permeability.

Incorporation of fibronectin into matrix decreases TNF-induced increase in endothelial monolayer permeability.

Plasma fibronectin, a dimeric adhesive protein in blood, incorporates into the subendothelial and interstitial matrix in the lung especially during vascular injury. Fibronectin in the matrix is believed to influence cell-cell interaction and endothelial cell adhesion to the collagen-rich extracellular matrix. We previously observed that addition of purified soluble human plasma fibronectin (hFn) to cultured pulmonary endothelial monolayers attenuates the increase in protein permeability of such monolayers exposed to tumor necrosis factor-alpha (TNF-alpha). In the current study, we determined the specificity of this permeability response to fibronectin by comparing hFn to two other purified adhesive proteins in human plasma, i.e., vitronectin (Vn) and fibrinogen (Fg). We also determined whether matrix incorporation was essential for this hFn-mediated protective response by comparing normal intact hFn to either hFn alkylated with N-ethylmaleimide (NEM) or to purified 160/180-kDa hFn fragments, since these alternate forms of fibronectin are believed to exhibit limited ability to incorporate into matrix. Calf pulmonary artery endothelial (CPAE) monolayers (3-4 days postseeding) were exposed to human recombinant TNF-alpha for 18 h at a medium concentration of 200 U/ml followed by assessment of protein permeability using transendothelial 125I-labeled albumin clearance. Dimeric hFn (600 micrograms/ml) significantly (P < 0.05) reduced the TNF-induced increase in endothelial monolayer permeability. Vn or Fg, added at equal molar concentrations to the hFn, were unable to attenuate endothelial permeability. Immunofluorescent analysis utilizing antibodies specific to either hFn, human Vn, or human Fg revealed incorporation of the exogenous hFn into the extracellular matrix, but no matrix incorporation of Vn or Fg. Both NEM-treated dimeric hFn as well as purified 160/180-kDa fragments of hFn, which cannot incorporate into the matrix, were also unable to prevent the TNF-induced increase in protein permeability. Thus the ability for soluble hFn to reduce the TNF-induced increase in lung endothelial monolayer permeability was specific and dependent on its incorporation into the extracellular matrix.

Immunofluorescent analysis of plasma fibronectin incorporation into the lung during acute inflammatory vascular injury.

Incorporation of plasma fibronectin into tissues is believed to influence endothelial cell-cell interaction, as well as endothelial cell adhesion to matrix. We used immunofluorescent microscopy coupled with tissue extraction of noncovalently incorporated fibronectin to delineate the time course for matrix incorporation of soluble plasma-derived fibronectin into the lung of sheep during postoperative bacteremia. Adult sheep were surgically prepared with both lung and peripheral lymph fistulas. Sheep were anesthetized 2 days following surgery and injected intravenously with a sublethal dose of live Pseudomonas aeruginosa, which consisted of 5 x 10(8) live organisms suspended in 0.9% saline. Bacterial infusion elicited a 300% increase in lung transvascular protein clearance but no increase in peripheral transvascular protein clearance. Purified dimeric human plasma fibronectin (hFn), used as an "immunologic marker," was then infused intravenously (100 mg/sheep) into two additional groups of sheep (nonbacteremic control group and bacteremic experimental group) and allowed to mix with the plasma pool of endogenous soluble sheep fibronectin (sFn). Incorporation of the plasma-derived hFn into the lung matrix and its distribution in relation to endogenous sheep fibronectin in the matrix was assessed by dual-label immunofluorescence using antibodies specific to either sFn or hFn. Human fibronectin from the vascular compartment codistributed with endogenous sheep fibronectin in the lung matrix. Moreover, its deposition into the lung was markedly increased in postoperative bacteremic sheep compared with nonbacteremic control sheep. Increased hFn deposition in the lung with bacteremia was clearly apparent within 2 h. The hFn deposited in the lung was nonextractable using a heparin-urea tissue extraction buffer, suggesting its rapid covalent cross-linking and incorporation into the lung matrix. Microscopic analysis of serial lung biopsies revealed focal areas of inflammation with an intense mononuclear infiltrate into the lungs by 2 h in the bacteremic sheep. Interstitial edema and vascular endothelial injury were observed by 4 h, with alveolar edema apparent over 6 to 8 h. Thus, postoperative bacteremia results in a rapid incorporation of plasma fibronectin into the lung matrix. This may be a physiologic mechanisms to stabilize the integrity of the lung vascular barrier.

Hepatic removal of 125I-DLT gelatin after burn injury: a model of soluble collagenous debris that interacts with plasma fibronectin.

The decline of plasma fibronectin after surgery, trauma, and burn, as well as during severe sepsis after injury, appears to limit hepatic Kupffer cell phagocytic activity. Intravenous infusion of gelatin-coated particles to simulate blood-borne particulate collagenous tissue debris in the circulation after injury also depletes plasma fibronectin. We used soluble gelatin conjugated with 125I-labeled dilactitol tyramine (DLT-gelatin) as a model of soluble collagenous tissue debris. We studied its blood clearance as well as organ localization in normal and postburn rats. Fibronectin-deficient plasma harvested early after burn exhibited limited ability to support in vitro phagocytic uptake of the gelatinized microparticles by Kupffer cells in liver tissue from normal rats. However, Kupffer cells in liver tissue from normal and postburn rats phagocytized the test particles at a normal rate when incubated in normal plasma. The DLT-gelatin ligand bound to fibronectin in a dose-dependent manner as verified by its capture with anti-fibronectin coated plastic wells when coincubated with purified fibronectin. By gel filtration chromatography, the binding of fibronectin with the DLT-gelatin ligand was readily detected, resulting in the formation of a high-molecular-weight complex. In normal animals the plasma clearance and liver localization of 125I-DLT-gelatin was competitively inhibited by infusion of excess nonradioactive gelatin. The blood clearance and liver localization of the soluble gelatin ligand were also impaired after burn injury during periods of fibronectin deficiency similarly to the pattern observed with gelatin-coated microparticles. By autoradiography, the cellular site for the uptake of the 125I-DLT-gelatin was primarily but not exclusively hepatic Kupffer cells; 125I-DLT-asialofetuin and 125I-DLT-ovalbumin were removed by hepatocytes and sinusoidal endothelial cells, respectively. Thus, gelatin conjugated with 125I-DLT can be used to simulate blood-borne soluble collagenous tissue debris after burn. It rapidly binds to plasma fibronectin before its hepatic Kupffer cell removal, and its blood clearance is markedly delayed after burn injury during periods of plasma fibronectin deficiency.

ED1-containing cellular fibronectin release into lung lymph during lung vascular injury with postoperative bacteremia.

Fibronectin (Fn) exists in both a soluble and insoluble form. Soluble Fn in plasma and lymph is an opsonic molecule that enhances phagocytic host defense. Insoluble Fn in the subendothelial and extracellular matrix is an adhesive molecule that mediates cell adhesion to substratum. The extracellular matrix of tissues such as the lung contains a mixture of both plasma-derived fibronectin (pFn) as well as locally synthesized cellular fibronectin (cFn). cFn is antigenically related to pFn, but cFn has extra domains (ED1 and ED2) that do not exist in liver synthesized pFn. The purpose of this study was to determine whether ED1-Fn was released into lung lymph before an increase in lung vascular permeability following postoperative bacteremia. Male sheep (n = 8) with surgically prepared lung lymph fistulae were infused intravenously with a sublethal dose (5 x 10(8)) of Pseudomonas aeruginosa 2 days following surgery. Lymph flow (QL), lymph-to-plasma (L/P) total protein ratio, lung protein clearance (QL x L/P), and hemodynamics were measured over 48 h following bacterial challenge. The lymph and plasma ED1-Fn concentrations were determined by enzyme-linked immunosorbent assay (ELISA) using a murine monoclonal antibody specific to the ED1 region of human cFn. There was a rapid rise of ED1-Fn flux in lung lymph which was evident 60 min after the start of bacterial infusion, resulting in a maximum three- to fourfold increase (P < 0.05) in this parameter. In contrast, the ED1-Fn concentration in plasma before bacterial infusion was less than lung lymph and it did not increase over the initial 6 h following bacterial infusion.(ABSTRACT TRUNCATED AT 250 WORDS)

Effect of fibronectin on permeability of normal and TNF-treated lung endothelial cell monolayers.

Fibronectin is found in a soluble form in plasma and lymph and in an insoluble form in the extracellular matrix. Plasma fibronectin can incorporate into the tissue pool of fibronectin where its adhesive properties may influence cell-cell interaction, cell adhesion to a collagenous matrix, and vascular integrity. Elevation of plasma fibronectin can attenuate the increase in lung vascular permeability in sheep during postoperative gram-negative bacteremia, and plasma fibronectin deficiency can magnify the increase in lung vascular permeability with postoperative sepsis. Using pulmonary endothelial monolayers, we determined if exogenous human plasma fibronectin (pFn) would influence the protein permeability of pulmonary endothelial monolayers as determined by transendothelial clearance (microliters/min) of 125I-albumin after they were exposed to human recombinant tumor necrosis factor-alpha. Treatment of endothelial monolayers with tumor necrosis factor (TNF) (200 U/ml) for 18 h resulted in a significant (P < 0.05) increase in protein permeability. Addition of intact purified human plasma fibronectin to normal confluent endothelial monolayers to yield a medium concentration of 300, 600, and 900 micrograms/ml for 18 h had no effect on baseline protein permeability. In contrast, whereas addition of lower amounts of human plasma fibronectin (300 micrograms/ml) did not attenuate the TNF-induced increase in monolayer permeability, the higher concentrations of 600 or 900 micrograms pFn/ml significantly decreased (P < 0.05) protein permeability. The ability of soluble plasma fibronectin to attenuate the TNF-induced increase in endothelial protein permeability required an incubation time of at least 2-3 h, perhaps due to a lag time required for its incorporation into the extracellular matrix.(ABSTRACT TRUNCATED AT 250 WORDS)

Rebound elevation of fibronectin after tissue injury and ischemia: role of fibronectin synthesis.

Plasma fibronectin (pFn) stimulates macrophage phagocytosis of tissue debris; pFn deposition in tissues may influence vascular integrity. Although the acute depletion of pFn after surgery and/or injury has been described, less attention has been given to the rebound hyperfibronectinemia presumably "triggered" by the early pFn depletion. Using a model that compartmentalized the site of tissue injury and thus attenuated the initial pFn depletion, we studied this rebound elevation of pFn in anesthetized rats (250-350 g) after the surgical trauma of groin dissection alone (sham group) or surgery coupled with 4 h of hindlimb ischemia (experimental group). Nonoperated control rats were also anesthetized. Shams had baseline (preoperative) 6-, 8-, and 22-h postoperative pFn levels of 573 +/- 61, 598 +/- 62, 695 +/- 57, and 929 +/- 87 micrograms/ml, respectively. In the surgery-ischemia group, pFn also elevated to 1,117 +/- 40 micrograms/ml at 22 h postsurgery. Nonoperated control rats (only anesthetized) had no elevation of pFn. Intravenous infusion of gelatin-coated lipid particles (50 mg/100 g) depleted pFn by 89.3% but was unable to prevent the rebound elevation of pFn. The blood clearance of 125I-labeled pFn was very similar in control, sham, and experimental rats. In contrast, pFn synthesis over the 22-h period was dramatically altered and equal to 2.12 +/- 0.16, 3.40 +/- 0.56, and 4.49 +/- 0.17 mg pFn synthesized/100 g body wt, in control, sham, and experimental rats respectively. Thus a rapid increase in pFn synthesis contributes to the rebound hyperfibronectinemia after sublethal surgical injury.(ABSTRACT TRUNCATED AT 250 WORDS)

Liver and spleen phagocytic depression after peripheral ischemia and reperfusion.

Liver and spleen phagocytic clearance of blood-borne microparticulate tissue debris and products of intravascular coagulation after trauma and surgical injury is an important mechanism to limit the deposition of debris in the pulmonary vascular bed. Plasma fibronectin (pFn) modulates this clearance process. We evaluated the effect of a localized peripheral ischemia and reperfusion injury on liver and spleen phagocytic function. Male rats (250 to 350 g) underwent 4 hours of tourniquet-induced bilateral hindlimb ischemia, followed by 18 hours of reperfusion after release of the tourniquet. Rats subjected to ether anesthesia alone or anesthesia followed by groin incision without ischemia were the control and sham groups, respectively. Reticuloendothelial (RE) phagocytic function was assessed at 15 minutes and 18 hours after the start of reperfusion by the in vivo liver and spleen removal of blood-borne iodine 125 (125I)-test microparticles, which were coated with gelatin (denatured collagen) to enhance their interaction with pFn. Liver and spleen particle uptake in control and sham rats was similar. In contrast, after 4 hours of ischemic injury with 15 minutes of reperfusion, we observed a 30% to 40% decrease (p less than 0.05) in liver and spleen particle uptake as compared with sham controls with partial restoration of this removal mechanism by 18 hours. This depression in liver and spleen phagocytic function was associated with a significant (p less than 0.05) increase in the deposition of the 125I-test particles in the lung. RE depression was not due to a deficiency of pFn; indeed, a marked elevation (588 +/- 12 micrograms/mL versus 1,083 +/- 40 micrograms/mL) of pFn was observed by immunoassay over the 18-hour reperfusion interval. Comparative bioassay of humoral (opsonic) versus cellular (Kupffer's cell) activity revealed that Kupffer's cells in livers from controls or ischemia-reperfusion rats exhibited normal phagocytic function when incubated in plasma harvested from either control or 4-hour ischemic rats. The opsonic activity of plasma harvested after ischemia and reperfusion was also more than adequate, consistent with the immunoassay analysis. Thus, the impaired liver and spleen clearance mechanism after peripheral ischemia and reperfusion injury did not appear to be due to either a macrophage cellular deficit or a lack of pFn. This clearance depression may be mediated by splanchnic malperfusion, which is known to develop after peripheral ischemia and reperfusion and associated soft tissue injury.

Leukocyte elastase-independent proteolysis of gelatin-bound fibronectin by inflammatory macrophages.

Fragmentation of lung matrix fibronectin by proteases released from activated phagocytic cells has been implicated in lung vascular injury. We examined whether denatured collagen (gelatin)-bound fibronectin can be degraded by peritoneal exudate mononuclear phagocytes harvested from rats 96 h after intraperitoneal casein injection. Microtiter plates were pretreated with gelatin and then supplemented with purified 125I rat plasma fibronectin, which readily bound to the gelatin. Stimulated inflammatory exudate cells were added and proteolysis of the bound fibronectin was studied by the release of [125I]fibronectin fragments into the media. Following 2 h of incubation, peritoneal exudate mononuclear macrophages stimulated with opsonized zymosan released three times more radiolabeled fibronectin into the medium as compared to background controls, and 1.5 times more radiolabeled fibronectin as compared to cells not stimulated with zymosan. Western blot analysis and autoradiography confirmed the presence of fragments of fibronectin in the culture medium. Some of these fragments were clearly derived from the radiolabeled matrix, but others that were not labeled were potentially released directly from the added stimulated macrophages. The release of radiolabeled fibronectin was inhibited by N-p-tosyl-L-lysine chloromethyl ketone (TLCK), a trypsin specific inhibitor, but not by methoxysuccinyl-alanine-alanine-proline-valine-chloromethyl-ketone (AAPVCK), a leukocyte elastase-specific inhibitor. These results suggest that fibronectin bound to denatured collagen is susceptible to leukocyte elastase-independent enzymatic degradation by stimulated inflammatory exudate mononuclear phagocytic cells. Such proteolysis may mimic a pathological process associated with lung vascular injury during the sequestration of activated macrophages in the lung microcirculation and interstitium.

Blood-borne fragments of fibronectin after thermal injury.

Fibronectin is an adhesive protein that can promote phagocytosis and endothelial cell adhesion. Plasma fibronectin declines following burn in animals and patients, potentially due to its complexing with circulating collagenous debris as well as its rapid binding to sites of tissue injury. Such depletion of fibronectin initiates an opsonic deficiency of the plasma. In view of the sensitivity of fibronectin to proteolytic enzymes, an additional factor that could contribute to the decrease of plasma opsonic activity after burn is the proteolytic fragmentation of fibronectin in the blood. In the current study, we determined if fibronectin fragments appear in the blood of anesthetized rats after a sublethal full-thickness skin burn of 15% to 16% of body surface. Plasma fibronectin concentration was quantified by enzyme-linked immunosorbent assay and the presence of fibronectin fragments in plasma was determined by immunoblot analysis. All blood was collected in an antiprotease mixture to yield final plasma concentrations of 0.15% EDTA, 3mmol/L phenylmethylsulfonyl fluoride, and 3 mmol/L iodoacetate to prevent degradation of fibronectin after sampling. Plasma fibronectin decreased 60% to 70% within 30 minutes post-burn, and this low level lasted for at least 4 hours. Within 30 minutes post-burn, two prominent fragments of fibronectin with a molecular weight of 110 +/- 2.2 kd and 122 +/- 3.3 Kd, respectively, were also detected in the plasma. Peak concentration of these fragments was detected at 60 minutes post-burn, but their level declined by 4 hours. By 4 hours, both bands appeared to resolve into doublets. To rule out the possibility that the fragments of fibronectin detected in the plasma were actually generated by coagulation enzymes activated at the site of peripheral blood sampling, rapid direct inferior vena cava sampling was performed, which also yield the presence of the fragments. Thus, fibronectin fragments exist in the plasma following thermal injury. Because fragments of fibronectin can compete with the intact fibronectin molecule with respect to its ability to stimulate macrophage phagocytosis, such fragments may contribute to altered systemic phagocytic host defense following thermal injury. Furthermore, because fibronectin peptides can compete with matrix fibronectin and impair adhesion of cultured endothelial cells, such circulating fragments may also influence the integrity of the vascular barrier.

Kinetics of plasma fibronectin: increased lung tissue incorporation after postoperative bacteremia.

Fibronectin is an adhesive glycoprotein that may influence the permeability of the vascular barrier. We compared the plasma disappearance of purified human fibronectin (hFn) and its incorporation into lung tissue in control nonbacteremic and bacteremic sheep after surgery to determine the influence of postoperative bacteremia on the plasma clearance and lung deposition of hFn. Lymph fistulas were surgically prepared 48 h before either saline or bacterial challenge to allow for collection of timed samples of plasma, lung lymph, and peripheral lymph. On the day of the study, the sheep were anesthetized and given either 5 X 10(8) Pseudomonas aeruginosa in saline or saline alone. In parallel, they were infused intravenously with 100 mg of hFn, and the hFn concentration in plasma, lymph, and tissue extracts was subsequently determined by enzyme-linked immunosorbent assay over an 8-h experimental interval. Localization of endogenous sheep fibronectin (sFn) and the injected hFn in serial lung tissue samples was determined by dual-label immunofluorescence. In nonbacteremic control sheep, plasma hFn levels declined with an initial rapid slope (t1/2 = 0.53 +/- 0.19 min), followed by a second, gradual slope (t1/2 = 21.7 +/- 2.5 h), whereas the concentration of hFn in lung lymph and peripheral lymph rose exponentially with time. In bacteremic sheep, the early plasma disappearance of hFn was similar, but the second phase of plasma clearance was faster (t1/2 = 7.4 +/- 0.3 h). Lung tissue from control and bacteremic sheep contained the same level of extractable hFn. However, tissue extraction followed by immunofluorescence microscopy indicated that lung tissue from bacteremic sheep contained more nonextractable hFn than lung tissue from nonbacteremic sheep. Thus postoperative bacteremia that elicits acute lung vascular injury will increase the plasma disappearance of hFn and its incorporation into the lung tissue.

Pulmonary and peripheral lymph protein clearance during bacteremia after surgery.

An increase in pulmonary transvascular protein clearance is seen in sheep following postoperative bacteremia. We evaluated whether this increased protein clearance is specific for the lung by comparing in sheep the effect of postoperative bacteremia on both pulmonary and prefemoral lymph protein clearance. Fibronectin, implicated as a factor influencing vascular permeability, was also measured. After intravenous infusion of a sublethal dose of 5 x 10(8) live Pseudomonas aeruginosa (n = 10), pulmonary lymph flow (QL) increased 146% (P less than 0.01), and the lung lymph-to-plasma (L/P) total protein concentration ratio increased 21% (P less than 0.05). This resulted in a 208% elevation (P less than 0.01) in lung protein clearance (LPC = QL x L/P). In contrast, peripheral lymph flow and peripheral protein clearance were not altered. After intravenous infusion of a lethal dose of 5 x 10(9) live Pseudomonas (n = 7), QL rose 243-348% (P less than 0.025), whereas the lung L/P remained at base line, resulting in a 240-358% increase in LPC (P less than 0.025). Again, peripheral lymph flow and protein clearance did not increase. Plasma fibronectin declined slightly after low-dose bacterial challenge but decreased (P less than 0.001) 40% by 4-5 h after high-dose bacterial challenge. After both low- and high-dose bacterial challenge, fibronectin in pulmonary lymph increased (P less than 0.05) relative to total protein. In contrast, no change in peripheral lymph fibronectin was seen. Thus, while postoperative bacteremia increased lung protein clearance, it did not increase peripheral protein clearance, suggesting specificity with regard to the microvascular response to bacteremia.

Plasma fibronectin therapy and lung protein clearance with bacteremia after surgery.

Plasma fibronectin modulates macrophage phagocytic function and can also incorporate into the insoluble tissue pool of fibronectin where it influences endothelial cell adhesion and tissue integrity. We studied the effect of postoperative bacteremia on lung protein clearance in relation to plasma fibronectin levels using the unanesthetized sheep lung lymph fistula model and the effect of infusion of purified human plasma fibronectin on lung protein clearance. Sheep received live Pseudomonas aeruginosa (5 X 10(8) iv) at a time of normal plasma fibronectin (590 +/- 37 micrograms/ml) or 5 days later at a time corresponding to elevation of plasma fibronectin (921 +/- 114 micrograms/ml). After the first bacterial challenge, there was a 22% decrease (P less than 0.05) in plasma fibronectin. Lung lymph flow (QL) initially increased 308% (P less than 0.05) by 2 h (0 h = 4.7 +/- 1.1 ml/h; 2 h = 14.4 +/- 3.5 ml/h), and the total protein lymph-to-plasma concentration ratio (L/P) declined. This was followed by a sustained second phase response over 3-12 h which was characterized by a 202-393% elevation in QL (P less than 0.05), an increase in the L/P ratio, and a 240-480% (P less than 0.05) increase in lung transvascular protein clearance (TVPC = QL X L/P). Sheep with elevated fibronectin levels also manifested the early (2 h) elevation in QL (P less than 0.05) coupled with a decline in L/P ratio after the second bacterial challenge, but the second-phase increase in TVPC was markedly attenuated. Intravenous infusion of 500 mg of human plasma fibronectin into normal sheep to elevate the fibronectin level comparable to that in the hyperfibronectinemic sheep also attenuated (P less than 0.05) the second-phase (3-12 h) increase in lung protein clearance with sepsis. Thus elevation of plasma fibronectin during postoperative Gram-negative bacteremia may protect the lung vascular barrier. This response may be mediated by either fibronectin's opsonic support of phagocytic function or its influence on lung endothelial cell adhesion.

Reversal of opsonic deficiency in surgical, trauma, and burn patients by infusion of purified human plasma fibronectin. Correlation with experimental observations.

Plasma fibronectin deficiency has been documented in critically ill surgical, trauma, and burn patients. Human plasma fibronectin was isolated by gelatin-Sepharose affinity chromatography and evaluated with respect to its opsonic activity following pasteurization, its in vivo clearance kinetics, and its short-term influence on cardiovascular hemodynamics in postoperative septic sheep. Six patients with low plasma fibronectin levels were also evaluated with respect to temporal changes of immunoreactive fibronectin and opsonic activity following infusion of fibronectin at a dose calculated to elevate the plasma fibronectin level to 400 micrograms/ml. With utilization of three different in vitro radioisotopic phagocytic assays, i.e., liver slice assay, peritoneal macrophage monolayer assay, and Kupffer cell monolayer assay, retention of opsonic activity by fibronectin following pasteurization was documented. The normal biphasic kinetics associated with plasma clearance of fibronectin were also not altered by pasteurization. In postoperative septic sheep with hemodynamic instability, intravenous infusion of 500 mg of purified human fibronectin initiated no abnormal hemodynamic response. Indeed, as compared with placebo, the infusion of fibronectin into the postoperative septic sheep resulted in a more stable systemic vascular resistance and pulmonary vascular resistance with a higher arterial pressure. It also elevated immunoreactive fibronectin levels (p less than 0.05) and increased opsonic activity (p less than 0.05). Surgical, trauma, and burn patients (ages 18 to 80 years) with low plasma fibronectin levels (160 to 236 micrograms/ml) manifested no disturbance in cardiovascular, respiratory, or hematologic parameters following fibronectin infusion (590 to 988 mg per patient), but did display an early increase of opsonic activity. This standardized, pasteurized, and opsonically active preparation of purified human plasma fibronectin (5.0 mg/ml after reconstitution) has utility for future randomized clinical trials in injured patients with sepsis.